INDEX

Explanations

interactions and dynamics between characters in tense situations

oai_token-act-pair · gpt-4o-mini Triggered by @bot

actions involving a person

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 surla

-0.60

 المعيارى

-0.56

himself

-0.54

setVerticalGroup

-0.53

 prêtres

-0.52

 himself

-0.50

tisseur

-0.48

 singur

-0.47

 mwenye

-0.46

IUrlHelper

-0.46

POSITIVE LOGITS

 themselves

0.74

 collectively

0.68

 själva

0.63

各自

0.60

themselves

0.59

 yourselves

0.57

 backgrounds

0.51

 their

0.51

 selves

0.50

 unanimously

0.49

Activations Density 0.074%

interactions and dynamics between characters in tense situations

actions involving a person

No Comments

No Known Activations